Risk Management for Binary Trading: Sizing and Limits

Your stake size decides whether you are still trading in three months. Your chart reading decides how well you do if you are. Most people spend all of their time on the second one and lose the account to the first.

That sounds like a slogan, so here is the arithmetic behind it. A trader with a real 57 percent win rate at an 85 percent payout makes money. Give that same trader a 20 percent stake size and a six loss run, which arrives roughly once in every 120 trades, and the account is down to 26 percent of where it started. Their edge is intact. Their account is not.

What does risk per trade mean when the whole stake is at risk?

In spot forex, risk and position size are separate. You can put $5,000 into a position and risk $50, because the stop loss defines the loss. In a fixed-time trade there is no stop loss. The stake is the loss. If you put $50 on the trade, you are risking $50.

This makes sizing simpler and much less forgiving. There is no clever entry that reduces the damage, no wide stop to survive a spike, no partial close. The number you type into the stake box is your risk, exactly, every time.

So the entire risk plan reduces to one question: what percentage of the balance goes into each trade?

How much should you risk per trade?

One to two percent. That is the answer, and the reason is in the table below rather than in tradition.

Recalculate it when the balance changes by a meaningful amount, not after every trade. Once a week is fine. What matters is that the number is decided before the session and written down somewhere you cannot quietly edit mid session, because the stake box is the easiest place in trading to lie to yourself.

Why a percent and not a fixed cash stake?

Because a fixed cash stake becomes a bigger risk exactly when you can least afford it. Take a $500 account with a flat $10 stake. That is 2 percent on day one. After a bad month at $250 it is 4 percent, and after a worse one at $150 it is 6.7 percent of everything you have left, on a single click.

Percent sizing fixes that automatically, and it works in both directions:

The gap is not enormous at 20 losses. The point is what happens after that. The flat stake account keeps risking a rising percentage until it reaches zero in a straight line. The percent account gets smaller stakes as it shrinks and mathematically cannot reach zero from losses alone, which buys you the one thing you need: time to notice and stop.

On the way up it works the same way in reverse. A 2 percent stake on a growing balance raises the cash stake for you, in step with the account, without you deciding to be braver on a Tuesday.

What losing runs do at each size

Losing runs are not a punishment for bad trading. They are what a 55 percent win rate looks like from the inside. At a 55 percent win rate the chance of losing any given trade is 45 percent, so:

Over 500 trades you should expect to meet a seven loss run somewhere, probably twice. That is not pessimism, it is what those numbers mean. Now put stake sizes against a ten loss run, compounding each loss on the reduced balance:

At 2 percent, a ten loss run is a bad fortnight you can trade out of. At 20 percent it is over. The account that needs plus 831 percent to get level is not coming back, and the trader knows it, which is when the sizing goes up again and the last of it disappears.

Turn the percent into cash before the session

Percentages are easy to bend in the moment. Cash numbers are not. Before you open the desk, write four lines somewhere you can see them:

• Stake per trade: $10.
• Maximum loss today: $30.
• Number of trades allowed: 6.
• MTG base and step if used: $3.33 and $6.67.

The reason to write the cash out is that "2 percent" is an abstraction your brain will happily reinterpret after two losses, while "$10" is a fact sitting next to the keyboard. Traders who type percentages into a calculator mid session end up rounding upward every single time, and they never round upward when they are winning.

Recalculate the four lines weekly, or when the balance moves more than about 20 percent. Not after each trade, because adjusting the stake downward after every loss and upward after every win is a staking plan you did not choose and cannot measure.

What win rate do you actually need?

Sizing only makes sense next to the break-even point, so get that number first.

Break-even win rate is 1 / (1 + p), where p is the payout as a decimal. At an 85 percent payout that is 1 / 1.85 = 54.1 percent. At 80 percent it is 55.6 percent. At 70 percent it is 58.8 percent. Payouts move around by pair, expiry and session, so read yours off the screen and do the division rather than trusting a number in an article.

Expectancy per trade is the other half of the same idea:

expectancy = (win rate x payout) minus (loss rate x 1)

At a 57 percent win rate and an 85 percent payout: (0.57 x 0.85) minus 0.43 = 0.4845 minus 0.43 = 0.0545. You make 5.45 percent of your stake per trade on average. Stake 2 percent of the balance and that is 0.109 percent of the account per trade. Small. It compounds, and it only compounds if the account survives, which brings us to the thing that kills more accounts than bad analysis.

Why does martingale blow accounts?

Because the ladder grows exponentially and the balance does not. Here is the doubling sequence in cash on a $1 base, with no rhetoric attached.

Read row four again. You have lost $7, you stake $8 to fix it, you win, and you are still 20 cents down on the sequence. That is the payout doing its work. In a casino, red pays 1 to 1 and doubling does recover the run. At an 85 percent payout, doubling only recovers for the first two steps. After that a winning recovery trade still leaves you behind.

The ladder that would actually recover

To recover the losses and still collect your original target, the next stake has to satisfy: stake x 0.85 = everything already lost plus the original target. Work it through from a $1 base with a target of $0.85:

The multiplier is about 2.18 per step, not 2. And look at the right hand column against a real account size. On a $100 balance, six losses in a row have taken $89.49, and the seventh step of the ladder requires a $106 stake you cannot place. The run that does that arrives roughly once in every 268 trades at a 55 percent win rate.

The part that kills people is the good part

A doubling ladder wins most of the time. That is not a defence of it, it is the mechanism. Six loss runs are rare, so the ladder produces a long string of small recoveries, and every one of them teaches the trader that it works. The account curve looks better than a fixed stake curve for weeks. Then one run arrives and takes the entire balance plus all of the accumulated profit.

That is the second line in the chart at the top of this article. It was ahead the whole way until it was not.

No staking plan can create an edge

This one is worth doing properly because it settles a lot of arguments. Take a one step MTG: stake a base amount, and if it loses, stake double once and then stop regardless of the result. The three outcomes, in units of the base stake at an 85 percent payout, are:

• Win the first: plus 0.85, with probability w.
• Lose then win: minus 1 plus 1.70 = plus 0.70, with probability (1 minus w) x w.
• Lose both: minus 1 minus 2 = minus 3, with probability (1 minus w) squared.

Set the total to zero and solve for w:

0.85w + 0.70w(1 minus w) minus 3(1 minus w) squared = 0

Expand it and you get 3.70w squared minus 7.55w + 3 = 0, which solves to w = 0.5405. That is exactly the same 54.05 percent break-even you get from single trades at an 85 percent payout. The staking plan changed nothing about where break-even sits. It only changed how lumpy the ride is.

This generalises. Expectancy per dollar staked is fixed by your win rate and the payout. Total result is that expectancy multiplied by total money staked. Rearranging when the money goes in cannot flip the sign. If your win rate is below 1 / (1 + p), every staking plan loses, and the aggressive ones just lose faster.

Then why does the REDBERRY desk allow one step of MTG?

Because one step is a bounded decision and a ladder is not. The desk rules say Live account rather than broker demo, 1 to 2 percent risk per signal, and at most one step of MTG, with the reasoning that the first candle is sometimes just an error candle on an otherwise valid read.

If you use it, size the pair, not the first trade. A one step MTG puts base plus double the base at risk, which is three times the base. So:

1. Decide the total you will risk on the signal. On a $500 account at 2 percent, that is $10.
2. Divide by three. Base stake $3.33, the single step $6.67.
3. That is the whole sequence. There is no third trade, no matter how certain the chart looks.

The failure mode is obvious and common: people take a $10 base and then a $20 step, which is 6 percent of the account on one read, and after two of those the day is down 12 percent. One step, sized inside the budget, or no MTG. Both are defensible. A ladder is not.

Daily loss limits: the rule that saves the week

Set two numbers before the session starts:

• A loss limit. Three losing trades, or 6 percent of the balance, whichever comes first.
• A trade cap. A maximum number of trades for the session, so a quiet market cannot be filled with clicks.

The reason the limit is small is that small holes are fillable. Down 6 percent needs plus 6.4 percent to get level, which is a normal week at 2 percent risk. Down 30 percent needs plus 42.9 percent, which is a project. Down 50 percent needs plus 100 percent, which is most of a year of good trading.

A loss limit is not about the money you save on the day. It is about never being in the right hand column.

There is a hard version of this built into the desk itself. If the Live balance drops below $10 the license ends on its own, and getting it back means depositing $50 so the balance is at $10 or more again. It is a product rule rather than a risk rule, but it points the same way: the account is the thing you are protecting, and the desk stops working when the account is gone. The seat and license mechanics are set out in REDBERRY license and seat rules.

Should you size up when you feel sure?

Test it before you believe it. Add one column to your log with a confidence score of 1, 2 or 3, filled in before the trade settles, then compare the win rate of your threes against your ones after fifty trades.

Most traders who run that test find the three groups sitting within a few points of each other, and some find their threes are worse, because high confidence tends to come from setups that look obvious, and obvious setups are where everyone else is positioned too. If your own log shows no relationship, confidence sizing is just variance with extra steps.

If the log does show a real gap, size the difference small. A 1 percent base with a 2 percent maximum on your best pattern is a defensible spread. A 1 percent base with a 10 percent hero trade is not, because a single hero loss undoes a week and the feeling of certainty is loudest right before the trades you most regret.

Weekly and monthly brakes

Daily limits stop a session. They do not stop a bad fortnight, which is where accounts actually die. Two more brakes, both boring:

• Weekly: three losing days, or minus 12 percent on the week, and the week is finished. Not reduced. Finished.
• Monthly: if the month closes red, next month starts at 1 percent per trade until the balance is back to where the month began.

These are not motivational rules, they are circuit breakers. The trader who ignores the daily limit on Monday, Tuesday and Wednesday is the trader who has an eight step ladder open on Thursday.

How do you recover from a drawdown?

Slowly, at the same stake size, or not at all. The recovery is arithmetic and the arithmetic is unsympathetic.

Say you are down 20 percent and you need plus 25 percent to get level. At a 57 percent win rate, an 85 percent payout and 2 percent per trade, you make about 0.109 percent of the balance per trade. Getting plus 25 percent from there takes roughly 205 trades. At 60 percent win rate it is about 100 trades. Either way it is weeks of ordinary trading, and there is no version of this where a bigger stake makes it shorter without also making the hole deeper.

What actually works:

1. Cut the size, not the standards. Drop to 1 percent while you rebuild. Lower stakes reduce the pressure that caused the hole.
2. Cut the trade count. Fewer, better setups. A drawdown is usually accompanied by a trade count that has quietly doubled.
3. Go back to one pair and one expiry. Complexity is expensive when your judgment is already bruised.
4. Read your own log before changing anything. If the losses cluster in one session or one expiry, the fix is a rule, not a new strategy.
5. Accept the timeline. Two hundred trades at 1 to 2 percent. Write that number down so the impatience has something to argue with.

What does not work, in every case for the same reason: raising the stake to speed it up. The stake rise increases both the expected gain and the variance, and the variance is what closes accounts. You are not trying to get level fast, you are trying to still be trading when you get level.

What if the account is small?

Small accounts have a real structural problem. If the smallest stake your platform allows is $1, then on a $50 balance that minimum is already 2 percent, and on a $30 balance it is over 3 percent. You cannot size properly below a certain balance, which means the sizing rule is doing nothing and the losing run does whatever it wants.

Two honest options. Trade the minimum stake and accept that your effective risk per trade is higher than you want, keeping the trade count very low. Or wait until the balance can carry the minimum stake at 2 percent, which for a $1 minimum means $50 and comfortably means $100. Note that unlocking a REDBERRY license needs a $100 Live deposit through the register link anyway, so the starting balance and the sizing floor line up reasonably well.

What is not an option is taking a $50 account and trading $5 stakes because the percentages look too slow. That is 10 percent per trade, and the table above already told you what a ten loss run does to it.

The one page risk plan

Written down, before the session, in this order:

1. Account balance today.
2. Stake per trade in cash: 1 to 2 percent of that balance.
3. Break-even win rate for the payout you are seeing: 1 / (1 + p).
4. Maximum trades for the session.
5. Loss limit: three losses or 6 percent, then the platform closes.
6. MTG: one step or none, sized so base plus step stays inside your per signal budget.
7. Where the log lives, and the columns it has.

Six lines and a location. The traders who survive their first year are usually not the best readers of price, they are the ones who wrote this down and did not renegotiate it at 11pm after a bad session. The negotiating part is a psychology problem, and it is covered in trading psychology and discipline.

Two final connections worth making. The clock you choose changes your win rate, so your expiry decision is also a risk decision, which is the whole argument in binary options expiry times explained. And the structural reason all of this is necessary, the fixed payoff and the full stake at risk, is laid out in what is binary trading. If you are working through the list of ways people break these rules in practice, common binary trading mistakes has the chart situations that cause each one.

REDBERRY publishes the risk language on the about page in the same plain terms: a non-profit informational desk, not a broker, no funds held, no trades placed, signals that are not 100 percent accurate. The sizing is yours. It is also the only part of this game you fully control.